Many third party viewers that support thumbnail views of documents or other sets of objects have a restriction of certain formats and size of the document or other object set. The thumbnails are also typically solely used for the purpose of navigating within a document or other object set. Maintaining the integrity of very large documents while allowing page modification or other operations, for example, is a big challenge. As a result, typical client side frameworks that support an infinite view of datasets (i.e., not limited by number of pages or other objects to be displayed, e.g., as thumbnails) typically only allow a user to view and navigation through them.
For a large document or other dataset, the viewer typically cannot maintain persistently any changes done by a user to the dataset, if permitted, as they get overwritten when fetched again from server. Current viewers typically do not support selection of items within infinite datasets; instead, knowledge of the selected state of a selected object typically gets lost during scrolling.